- Email: [email protected]
Breastfeeding and Brain Development
BREASTFEEDING 2001, PART I: THE EVIDENCE FOR BREASTFEEDING
0031-3955/01 $15.00 + .OO
BREASTFEEDING AND BRAIN DEVELOPMENT Ann Reynolds, MD
Although children who are breastfed have better neurodevelopmental outcomes, whether this is a biological or nutritional effect, an environmental effect, a genetic effect, or some combination of these factors is unclear. Because a randomized trial comparing breastmilk with formula is impossible, investigators must rely on observational studies, which creates methodologic difficulties. Potential confounding variables must be identified and addressed. Women who choose to breastfeed tend to have attributes, such as a higher level of education, that are associated with better developmental outcome12;however, evidence also supports a biological or nutritional role. Docosahexaenoic acid (DHA), a longchain polyunsaturated fatty acid present in large quantities in the brain and retina, is present in human milk but not in commercial formula. The role of human milk in neurodevelopmental outcome, with emphasis on visual and cognitive outcome in term and preterm infants, is reviewed. POLYUNSATURATED FAlTY ACIDS IN THE BRAIN
The role of polyunsaturated fatty acids in the function of the brain and retina is an area of active research. Although DHA and arachidonic acid are not present in infant formula, they can be synthesized from their essential fatty acid precursors a-linolenic acid (ALA) and linoleic acid, which are present in infant formula; however, an extra step seems to exist in the synthesis of DHA that slows down the conversion of ALA ~
From the Child Development Unit, The Children's Hospital, University of Colorado Health Sciences Center, Denver, Colorado PEDIATRIC CLINICS OF NORTH AMERICA VOLUME 48 NUMBER 1 FEBRUARY 2001
159
160
REYNOLDS
to DHA." This step is potentiallyProblematic in infants who are rapidly incorporating DHA into their CNS. Without a dietary source, infants must synthesize large amounts of DHA and arachidonic acid to keep up with the needs of the developing brain. The most rapid accretion of DHA and arachidonic acid in the human brain occurs during the last trimester of pregnancys when these long-chain polyunsaturated fatty acids normally would be provided by the placenta." Rapid accretion of DHA during the last 3 months of gestation would make premature infants vulnerable to a deficiency of DHA, but evidence shows that dietary DHA also may affect term infants. In a study35of infants who died of sudden infant death syndrome, DHA concentrations in the cortex were initially similar in newborns whether they received human milk or infant formula (7%); however, during the first 48 weeks of life, DHA concentrations in the cortex increased in infants who were breastfed to lo%, whereas concentrations in formula-fed infants remained the same. In another study of term infants who died of SIDS,'6 a postmortem examination found greater parietal cortex DHA in breastfed infants (9.7%) than in formula-fed infants (7.6%) at 9 weeks of age (P < 0.02). Infants who receive DHA in their diets seem to incorporate more DHA into their brains; however, whether any functional advantage to having higher levels of DHA in the brain exists is unclear. VISUAL FUNCTION IN BREASTFED INFANTS
Photoreceptor membranes in the retina have the highest concentration of DHA in the body. Photoreceptor cells undergo rapid maturation in the last trimester of pregnancy and the first 4 to 6 months postpartum. Consequently, investigators speculate that dietary availability of DHA should correlate with visual function. Neuringer et ala found functional visual deficits in rhesus monkeys who were deficient in DHA and ALA. Many studies also have been done in human infants to determine the effects of human milk or dietary fatty acids on visual function, the results of which have been incon~istent.'~ A positive effect of dietary DHA on visual outcome is more likely to occur in premature infants than in term infant^.^, 19,34, 57 Uauy et a157compared breastfed premature infants with three groups of formula-fed infants. The formula-fed infants were randomly assigned to formula deficient in 0-3 fatty acids, standard formula, or formula supplemented with marine oil that contains the w3 long-chain polyunsaturated fatty acids DHA and eicosapentaenoic acid. Thresholds for rod electroretinograms (ERGs) were significantly higher, and amplitudes of rod ERGs were significantly lower for the deficient group compared with the marine-oil-supplemented and breastfed groups. Results for the infants supplemented with marine oil were similar to those of breastfed infants. The infants fed standard formula had intermediate results that did not differ sigruficantly from the deficient groups or from the supplemented and breastfed gr0ups.5~The deficient group was deficient not only in DHA but also in its precursor, ALA. Carlson et a17 studied 67 premature infants randomized to stan-
BREASTFEEDING AND BRAIN DEVELOPMENT
161
dard formula or marine-oil-supplemented formula. Visual acuity was measured by preferential looking, a behavioral measure that can be used for infants. Infants prefer to look at a pattern instead of a homogeneous background; if an infant prefers a stimulus card with a checkerboard to a blank card, then one can assume that the squares in the checkerboard are larger than the infant’s threshold of resolution.41At 2 and 4 months of age, visual acuity measured by preferential looking was better in the marine-oil-supplemented group. No significant difference was found between groups at 6.5, 9.0, and 12.0 months. Visual acuity seems to ”catch up” in infants fed standard formula. It may have been more difficult to find a difference in the older age groups because visual acuity changes quickly in the first 6 months of life and then plateau^.^ In the same group at 9 and 12 months, Carlson6 reported increased discrete looks to faces during the novel portion of testing in the marine-oilsupplemented infants, which may indicate a “higher order of mental function.” Birch et a13also found better stereo acuity and letter matching in breastfed infants at 36 months of age. Most studies that have evaluated visual function did not adjust for socioeconomic status (SES), but a study by Makrides et a P showed that SES seemed to be associated with visual function. Another problem in interpreting these studies is that many of them use different oils for supplementing formula, many of which contain different types and.amounts of fatty acids, so comparison of the studies is difficult. Many of the studies also have small sample sizes that may limit their power.37The effect of breastfeeding on visual function is still unclear, especially in term infants. In preterm infants, there may be enhanced maturation of the visual system, which may affect the quantity and quality of early stimuli. Also, breastfeeding may affect other aspects of visual function that are seDarate from visual acuity. Future &dies should include larger sampfe sizes and adjust for SES. DEVELOPMENTAL AND COGNITIVE OUTCOME OF BREASTFED INFANTS The effect of breastfeeding on later neurodevelopmental outcome is controversial. Because physicians must rely on observational studies, unforeseen biases may exist in the samples and may affect the validity of the results. Confounding variables also make the results difficult to interpret. Potential confounding variables include: Parental or environmental variables Maternal and paternal IQ (as it relates to genotype and environment) Maternal and paternal education level SES Maternal age Home environment Mother-infant interactions and bonding
Hormonal influences of lactation on mother Maternal stress Maternal depression Maternal attitudes regarding health practices Birth order Child variables Neonatal factors that might affect decision to breastfeed, such as illness or feeding dysfunction Birth weight Intrauterine growth retardation Gestational age Abnormal neurologic examination Abnormal neuroimaging Bronchopulmonary dysplasia Meningitis or sepsis Skin-to-skin contact in premature infants Although this list is long, it may be incomplete. Women who breastfeed have different attributes from those who choose not to breastfeed. Women who choose to breastfeed tend to be older, have more education, 5o These factors are associated with a better and have higher developmental outcome for Lucas et-a132noted that women who were married, primiparous, delivered by cesarean section, and had a boy were most likely to initiate breastfeeding in preterm Other differences may be more difficult to measure among mothers who choose to breastfeed and mothers who do not.= In a study9of mothers of low SES, mothers who chose to breastfeed had better vocabulary, were older, and were less likely to smoke cigarettes. The mothers who breastfed also were found to be less authoritarian, and they provided a more stimulating home environment. Another study"" found that mothers who breastfeed may be more likely to practice health-enhancing behaviors, such as not smoking, eating five or more fruits and vegetables per day, and visiting the dentist yearly. These health-enhancing behaviors may have multiple implications, including a better nutritional state during pregnancy. In addition to the prelactation differences among women who choose to breastfeed and mothers who do not, breastfeeding may affect mothers, infants, or both. Breastfeeding is a time of mother-infant bonding. Early interactions between mothers and infants seem to help to determine the type of relationship that infants will have with their environments over time.12.22 The hormonal influences of breastfeeding also may affect maternal behavior. In rats, oxytocin has been found to trigger maternal behavior in estrogen-primed In humans, it is more difficult to differentiate learned from hormonally driven behavior. In rats and sheep, hormonal influences on maternal behavior seem to wane after delivery. Nonhormonal, or psychological, influences take over as rats learn the odor or call of their pups.49Even if hormonal influences are present early on in humans, they may not be critical in
BREASTFEEDING AND BRAIN DEVELOPMENT
163
influencing maternal behavior. Maternal stress and depression also can affect infant development?, 56 Kavanaugh et aIz6reported emotional rewards in mothers who provided breastmilk for their premature infants. This greater sense of well-being in mothers also may affect infants. Many of these variables could be addressed by evaluating environmental stimulation and mother-hild interactions. Few studies have attempted to do that in a systematic way.E,39 Several studies have evaluated developmental or cognitive outcome in breastfed infants.* A meta-analysis2 was published in 1999. To be included in the meta-analysis, a study had to compare subjects who were predominantly breastfed with subjects who were predominantly formula-fed and use an outcome measure that was a ”widely applied test of cognitive development.” Twenty studies were used, but only 11 provided usable data with adjusted and unadjusted scores. All of the studies included in the meta-analysis adjusted for at least five confounding variables. The analysis combined studies of term and preterm infants. After adjusting for possible confounding variables, an average 3.2point higher ”cognitive development score” was found among breastfed infants. Figure 1 shows the covariate-adjusted mean differences, with confidence intervals, for the studies included in the meta-analysis. The advantage was seen early on (6-23 mo) and continued throughout childhood to 10 to 15 years of age? If the cognitive advantage of breastfeeding were entirely caused by environmental factors, then the difference in developmental scores would be expected to increase with age. Developmental evaluation tools primarily evaluate sensorimotor skills in the first 2 or 3 years of life and therefore are less affected by the environment. As children get older, the tools incorporate more language, which is heavily influenced by the The meta-analysis also found that duration of breastfeeding correlated with developmental and cognitive outcome. No significant difference was found when duration of breastfeeding was 4 to 7 weeks. The mean difference increased as the duration of breastfeeding increased. The difference was 2.91 among infants who were breastfed for 28 weeks or more.2As with studies of visual function, a bigger advantage is conveyed by providing human milk to premature infants compared with term infants. In the metaanalysis, low birth weight infants had a 5.18-point advantage if breastfed compared with a 2.66-point advantage among term infants.2At least one of the studies in the analysis had a major methodologic concern because breastfeeding was determined by record review of charts that were not always complete.18Two of the reports13,24 were letters to the editor in response to another article and so were not peer reviewed; however, one letter reported data from a study that was published laterE and is discussed in this article. Table 1 summarizes studies included in the meta-analysis. Three studies evaluating outcome in term infants fed breastmilk that were published after the meta-analysis are discussed here.”, E, 36 *References 1, 13, 17, 24, 27, 43,4548, 52, 54, and 55.
164
REYNOLDS
3.40
Rogers
229
Fergusson et al
5.61 Ounsted et al
4.36
Morley et al
4.72
Morrow-Tlucak et al
5.02
Doyle et al
Jacobson and Jacobson
1.71 8.30
Lucas et al
2.70
Rogan and Gladen
4.36
Ternboury et al
3.70
Florey et al
3.16
1
All studies - 4 - 2
0
2
4
6
8
1 0 1 2 1 4
Figure 1. Effect of breastfeedingversus formula feeding on cognitive developmentalscore: covariate-adjustedmean differences with confidence intervals. (from Anderson JA, Johnstone BM, Remley DT: Am J Clin Nutr 70:525-535,1999; with permission.)
Jacobson et a P reported the cognitive outcomes of 280 children at 4 and 11years of age. A sigdcant association was found between the duration of breastfeeding and IQ at 4 and 11 years of age. The association was weakened but remained significant after adjusting for SES and maternal education level. The association was not statistically significant after additional adjustment for maternal IQ and the score on the Home Observation for Measurement of the Environment (HOME)? The Peabody Picture Vocabulary Test-Revised14 was used to measure IQ. It is not a measure of IQ but of receptive vocabulary. It was used because it correlates with the Wechsler Adult Intelligence Scale (median’ r = 0.72). The HOME inventory combines observation and interview of the primary caregiver in the child’s home. The HOME inventory evaluates opportunity for stimulation and quality and amount of maternal interact i ~ nThe . ~ HOME inventory is ”sigzufrcantly and independently related to children’s cognitive development” even when mothers’ IQ, educational achievement, and SES are adjusted accordingly.2OA possible concern with this studyz is that the cohort was derived from a study of
BREASTFEEDING AND BRAIN DEVELOPMENT
165
Table 1. SUMMARY OF STUDIES USED IN META-ANALYSIS
Study
No. Patlents
Rogers"
2424
Fergusson"
1037
ounsteda
242
Morley"
771
Morrow-Tl~cak~~ 229
Doyle13
96
Jacobsonz4
323
Lucas31
283
~0gan47
855
Tembow Florey'S
229 592
Tests Used to Evaluate Outcome
Additional Information
Covarlate Adlusted Mean Difference
8 y: picture intelligence; 1946 cohort, tests 3.40' 15 y: nonverbal not adequately ability, sentence described completion 2.09' Controlled for 3 y: Peabody Picture Vocabulary Test; many things, 5 y: Stanford Binet; including home 7 y: WE€-R environment 7.5 y: British Abilities All mothers had 5.61* hypertension Scale (very large confidence interval) 18 mo: Bayley, Neuroimaging 4.36* Developmental not addressed, Profile II LBW 4.72' 6,18, and 24 mo: Difference only significant at 12 Bayley and 24 mo, controlled for home environment, low SES group 5.02' 2 y: Bayley; 5 y: WPPSI; Letter, LBW 8 y: WISC 1.71 4 y: McCarthy Letter, same cohort as Jacobson (1999) 7.5-8.0 y: WISC No discussion of 8.30' neuroimaging (Anglicized revised edition) or neurological status, LBW 2.70' 6,12,18,24mo: Bayley; Not always blinded, PCB 3,4,5 y: McCarthy cohort 18-29 mo: Bayley 4.36 18 mo: Bayley Breastfeeding 3.70' determined by retrospective record review
'Covariate adjusted mean differrnce is significant. WISC-R = Wechsler Intelligence Scale for C h i l b (Revised); Bayley = Bayley Scales of Infant Development; McCarthy = McCarthy Scales of Children's Abilities;LBW = Low birth weight; WPF'SI = Wechsler Preschool and Primary Scale of Jntelligence; SES = socioeconomicstatus; PCB = polychlorinated biphenyls.
long-term effects of polychlorinated biphenyls. To participate in the study, mothers had to have a history of consuming at least 11.8 kg of fish from Lake Michigan over a 6-year peri0d.2~This requirement may have caused some bias. The authors note that prenatal exposure to polychlorinated biphenyls can cause subtle cognitive deficits but exposure by human milk should not. Also, all of the potential confounding variables may be highly correlated with each other, which may create multicolinearity, making the coefficients obtained in a hierarchical regression possibly misleading.'O When maternal IQ and HOME score are added into the multiple-regression analysis, breastfeeding, SES, and education level are insignificant. The authors do not address colinearity; thus, to what extent maternal IQ and HOME score are related to breastfeeding is unknown. This question could be answered by looking at zero-order correlation. A stepwise, multiple regression adding in first HOME score then maternal IQ, SES, education level, and finally breastfeeding could be conducted. Then the extent to which breastfeeding added predictive value to the dependent outcome measures could be determined. Unfortunately, the tables in the article are not displayed in a traditional hierarchical manner that would have detailed the amount of variance in scores accounted for by each block of variables detailing incremental predictions for each block. This is a perfect example of how difficult it is to interpret data in a nonrandomized, observational study. This is especially true when confounding variables are highly correlated with each other. In another study initially designed to follow infants exposed to chloride-deficient soy formula between 1978 and 1979,36 outcome at 9 and 10 years of age was assessed. This study also showed no cognitive advantage for breastfed infants at 9 and 10 years of age after scores were adjusted for maternal education level, paternal education level, and household annual income. In this study, mothers were asked to retrospectively report the duration of breastfeeding at the 9-year or 10-year well-child visit.% This method is concerning because one study found that only 54% of mothers could correctly identify within 1 month when they had switched from human milk to formula when their children were 2 to 4 years of age.I5 The third study is a population-based study from a cohort in New Zealand who were followed up until 18 years of age. Breastfeeding status was determined prospectively. In that study, a 0.11 to 0.30 standard deviation greater score (correlates with 1.54.5 points) was found on IQ and achievement measures among individuals who had been breastfed for more than 8 months even after adjustment for confounding factors, including maternal age, maternal education level, SES, family income, and an informal assessment of the en~ironment.2~ All of these studies evaluated long-term outcome. Any biological effect may be picked up more easily early on, before environmental factors become predominate, but clinically, long-term outcome would be the endpoint. In term infants, the effect of breastfeeding on later cognitive outcome seems to be small or insignificant after adjusting for maternal and environmental
-
BREASTFEEDING AND BRAIN DEVELOPMENT
167
factors. Additional studies should focus on lower SES groups because one study found a 4.7-point advantage in lower SES children even after adjusting for the HOME score, parental authoritarianism, maternal Peabody Picture Vocabulary Test-Revised, smoking habits, and marital ~fatus.3~ Future studies might also focus on more specific areas of cognition and neurologic function. Neuropsychological testing might be able to detect more significant differences in areas such as memory, attention, and fine motor skills. PREMATURE INFANTS
Premature infants are discussed separately because their nutritional needs are different from those of term infants and because the effects of breastfeeding seem to be greater in premature infants. Most of the studies published on cognitive outcome in premature infants who are fed human milk present data from the same cohort of infants in the United Kingdom. The preterm infants ( 4 8 5 0 g) were followed prospectively as part of a study to determine the optimal type of nutrition in children who were not breastfed or whose mothers did not produce adequate breast milk. Comparisons were made between preterm formula and unfiltered donor breast milk, and between preterm formula and term formula.30,31, 33, 38 In this cohort, Lucas et aI3l compared preterm infants who received their mothers’ breast milk during their hospital course with infants who did not. The comparison group received preterm formula (51%),mature pasteurized donor drip milk without fortification (31%), or term formula (18%). Infants in the group fed their mothers’ milk were supplemented with preterm formula, donor breast milk, or term formula in the same ratios. Lucas et a131found a 10.9-point advantage in IQ measured by the Wechsler Intelligence Scale for Children (Revised) (WISC-R) at 7.5 to 8 years of age. After controlling for SES and maternal education level, an 8.3-point advantage remained. In an attempt to further control for maternal differences associated with choosing to breastfeed, the investigators compared those who chose to provide breast milk but were unable to with those who never intended to provide breast milk. The mean IQ of children whose mothers intended to provide breast milk but could not (94.8) was similar to the mean IQ of children whose mothers never intended to breastfeed (92.8). They were both sigruficantly different from the group who received their mothers’ milk (103.7). While in the nursery, neonates received breast milk principally by nasogastric tube. To control for the effects of suckling, Lucas et a131 excluded infants from the analysis who would be breastfed after discharge. The breast milk advantage remained but decreased slightly to 7.5 points. Because many infants who received their own mothers’ breast milk also required supplementation, Lucas et a131 compared the amount of mother’s milk received with IQ. A significant dose-response curve was found. Comparisons of the group who received mother’s milk with the group that did not showed no significant
168
REYNOLDS
difference in birth weight, gestational age, days in the study, days to enteral feeds, and the percentage Ventilated for more than 5 days. More boys were included in the group that received mother's milk (55% versus 42%) and, as expected, SES and maternal education level were higher in the group that provided breast milk. These data are compelling but should be interpreted with caution. The study by Lucas et aP' has two potential problems. The heterogeneity of feeding type in the comparison group makes difficult the generalization of these findings. Term formula is inadequate to meet the nutritional needs of premature infants, so how much its use affected the overall results is unclear. It would be helpful to see the individual relationships between the mothers' milk group and those fed with preterm formula or with those fed donor breastmilk. A significant potential problem with the study is that no mention was made of rates of intrauterine growth retardation, intraventricular hemorrhage, periventricular leukomalacia, ventricular dilitation, abnormal neurologic status, or severe infection?l All of these factors could have a sigruhcant effect on developmental outcome.4,40, 53 Some children in the cohort seemed to have intraventricular hemorrhage and cerebral palsy because those issues are addressed in other p~blications~~~ m; however, they are addressed as they apply to preterm formula compared with term formula or donor breast milk, not as they apply to mothers' milk compared with other feeding strategies. Additional study is needed to determine the role of breast milk in the cognitive outcome of premature'infants. Future studies should include more complete information regarding confounding variables, such as intraventricularhemorrhage, periventricular leukomalacia, Cerebral palsy, other neurologic dysfunction, maternal intelligence, and a measure of environmental stimulation. Maternal stress and depression and mother-infant bonding also would be important to address in preterm infants. Despite these concerns at the least, the type of enteral feeding should be considered a potential confounding variable in neurodevelopmental outcome studies in premature infants. SUMMARY
Although biochemical evidence seems to support the fact that more DHA is incorporated into the brain of breastfed infants compared with formula-fed infants, whether the levels of DHA in the brain are clinically significant is unclear. Because randomized trials cannot be done, this issue is difficult to study. The effects of breastfeeding on developmental outcome in term infants seems to be small or insignificant. For otherwise healthy children the potential differences are not clinically relevant; however, these small differences distributed over an entire population might have a significant effect on society. Although significant methodologic concerns exist, the effects of breastfeeding on preterm infants may be greater than those for term infants. Extremely low birth weight, premature infants (<750-1000 g) have been found to have IQs that are
BREASTFEEDING AND BRAIN DEVELOPMENT
169
13 points lower than term controls and a 50% to 60% risk for requiring special-education services when they are in school.2*, 51 In these infants, small improvements in IQ and neurologic function could have a much greater effect. Further study of neurodevelopmental outcome in premature infants fed breast milk compared with those fed preterm formula are indicated. This information should not change the practice of encouraging breastfeeding of term and preterm infants because other advantages to breastfeeding exist. References 1. Agostoni C, Trojan S, Bellu R, et a1 Developmental quotient at 24 months and fatty acid composition of diet in early infancy: A follow up study. Arch Dis Child 76:421424,1997 2. Anderson JW, Johnstone BM, Remley DT: Breastfeeding and cognitive development: A meta-analysis. Am J Clin Nutr 70525-535, 1999 3. Birch E, Birch D, Hoffman D, et a1 Breastfeeding and optimal visual development. J Pediatr Ophthalmol Strabismus 3033-38,1993 4. Brazy JE, Eckerman CO, Oehler JM, et a1 Nursery neurobiological risk score: Important factors in predicting outcome in very low birth weight infants. J Pediatr 118:783-792, 1991 5. Caldwell BM, Bradley RH: Home Observation for Measurement of the Environment. Little Rock, AR, University of Arkansas Press, 1979 6. Carlson S E Lipid requirements of very-low-birth-weight infants for optimal growth and development. In Dobbing J, Benson JD (eds): Lipids, Learning and the Brain: Fats in Infant Formulas. Report of the 103rd Ross Conference on Pediatric Research. Columbus, OH, Ross Laboratories, 1993 7. Carlson SE, Werkman SH, Rhodes PG, et a1 Visual-acuity development in healthy preterm infants: effect of marine41 supplementation. Am J Clin Nutr 58:35-42,1993 8. Clandinin MT, Chappell JE,Leong S, et al: Intrauterine fatty acid accretion rates in human brain: Implications for fatty acid requirements. Early Hum Dev 4:121-129,1980 9. Cogill SR, Caplan HL, Alexandra H, et al: Impact of maternal postnatal depression on cognitive development of young children. BMJ 292:1165-1167,1986 10. Cohen J, Cohen P: Multiple regression and correlation: Two or more independent variables. In Cohen J, Cohen P (eds): Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences. Hillsdale, NJ, Lawrence Erlbaum Associates, 1975 11. Crawford MA: Placental delivery of arachidonic and docosahexaenoic acids: Implications for the lipid nutrition of preterm infants. Am J Clin Nutr 71(supp1):275-284,2000 12. De Andraca I, Uauy R Breastfeeding for optimal mental development. The Alpha and the Omega in human milk. World Rev Nutr Diet 781-27,1995 13. Doyle LW, Rickards AL, Kelly EA, et a1 Breastfeeding and intelligence [letter]. Lancet 339744-745,1992 14. Dunn LM Peabody Picture Vocabulary Test-Revised. Circle Pines, MN, American Guidance Services, 1987 15. Eaton-Evans J, Dugdale AE: Recall by mothers of the birth weights and feeding of their children [abstract]. Human Nutr Appl Nutr 40371-175, 1986 16. Farfuharson J, Cockburn F, Patrick WA, et al: Infant cerebral cortex phospholipid fattyacid composition and diet. Lancet 340810-813,1992 17. Fergusson DM, Beautrais AL, Silva PA: Breastfeeding and cognitive development in the first seven years of life. Soc Sci Med 1617051708,1982 18. Florey CdV, Leech AM, Blackhall A Infant feeding and mental and motor development at 18 months of age in first born singletons. Int J Epidemiol24(suppl):21-26, 1995 19. Gibson RA, Makrides M Polyunsaturated fatty acids and infant visual development: A critical appraisal of randomized trials. Lipids 34A79-184, 1999
20. Gottfried AW. Home environment and early cognitive development:Integration, metaanalyses, and conclusions. In Gottfried AW (ed): Home Environment and Early Cognitive Development, Longitudinal Research. Orlando, Academic Press, 1984 21. Hack M, Taylor G, Klein N, et al: School-age outcomes in children with birth weights under 750 grams. N Engl J Med 331:753-759,1994 22. Hofer MA: Early relationships as regulators of infant physiology and behavior. Acta Paediatr 397(suppl):9-18,1994 23. Horwood LJ, Fergusson DM: Breastfeeding and later cognitive and academic outcomes. Pediatrics 101:91-97, 1998 24. Jacobson SW, Jacobson JL: Breastfeeding and intelligence [letter]. Lancet 339926, 1992 25. Jacobson SW, Chiodo LM, Jacobson J L Breastfeeding effects on intelligence in 4 and 11-year old children. Pediatrics 10371, 1999 26. Kavanaugh K, Meier P, Zimmerman B, et a1 The rewards outweigh the efforts: Breastfeeding outcomes for mothers of preterm infants. J Hum Lact 13:15-21,1997 27. Lanting CI, Fidler V, Huisman M, et al: Neurological differences between 9-year-old children fed breast-milk or formula-milk as babies. Lancet W1319-1322, 1994 28. Lawrence RA, Lawrence RM: Psychologic impact of breastfeeding. In Breastfeeding: A Guide for the Medical Profession, ed 5. St. Louis, Mosby, 1999 29. Lucas A, Morley R, Cole TJ: Randomised trial of early diet in preterm babies and later intelligence quotient. BMJ 3171481-1487, 1998 30. Lucas A, Morley R, Cole TJ, et a1 A randomised multicentre study of human milk versus formula and later development in preterm infants. Arch Dis Child 70F141F146,1994 31. Lucas A, Morley R, Cole TJ, et a1 Breast milk and subsequent intelligence quotient in children born preterm. Lancet 339261-264,1992 32. Lucas A, Cole TJ, Morley R, et a1 Factors associated with maternal choice to provide breastmilk for low birth weight infants. Arch Dis Child 63:48-52,1988 33. Lucas A, Gore SM, Cole TJ, et a1 Multicentre tial on feeding low birth weight infants: Effects of diet on early growth. Arch Dis Child 59:722-730, 1984 34. Makrides M, Neumann MA, Jeffrey B, et al: A randomized trial of different ratios of linoleic to a-linolenic acid in the diet of term infants: effects on visual function and growth. Am J Clin Nutr 71:120-129, 2000 35. Makrides M, Neumann MA, Byard RW, et a1 Fatty acid composition of brain, retina, and erythrocytes in breast- and formula-fed infants. Am J Clin Nutr 60189-194,1994 36. Malloy MH, Berendes H Does breast-feeding influence intelligence quotients at 9 and 10 years of age? Early Hum Dev 50:209-217,1998 37. Morley R Nutrition and cognitive development. Nutrition 14752-754,1998 38. Morley R, Cole TJ, Powell R, et al: Mother’s choice to provide breast milk and developmental outcome. Arch Dis Child 63:1382-1385,1988 39. Morrow-Tlucak M, Haude RH, et al: Breastfeeding and cognitive development in the first 2 years of life. SOCSci Med 26635-639, 1988 40. Msall ME, Buck GM, Rogers BT, et al: Risk factors for major neurodevelopmental impairments and need for special education resources in extremely premature infants. J Pediatr 119:606414, 1991 41. Neuringer M Infant vision and retinal function in studies of dietary long-chain polyunsaturated fatty acids: Methods, results, and implications. Am J Clin Nutr 71(suppl):256-267,2000 42. Neuringer M, Connor WE, Lin DS, et ak Biochemical and functional effects of prenatal and postnatal omega 3 fatty acid deficiency on retina and brain in rhesus monkeys. Proc Natl Acad Sci U S A 83:4021-4025,1986 43. Ounsted M, Moar V, Cockbum J, et al: Factors associated with intellectual ability of children born to women with high risk pregnancies. BMJ 2881038-1041,1984 44. Pesaa JA, Shelton MN: Health-enhancing behaviors correlated with breastfeeding among a national sample of mothers. Public Health Nurs 16:120-124,1999 45. Pollock J I Long-term associations with infant feeding in a clinically advantaged population of babies. Dev Med Child Neurol x429-440, 1994 46. Richards M, Wadsworth M, Rahimi-Foroushani A, et ak Infant nutrition and cognitive
BREASTFEEDING AND BRAIN DEVELOPMENT
171
development in the first offspring of a national UK birth cohort. Dev Med Child Neur 40:16%167,1998 47. Rogan WJ, Gladen BC: Breast-feeding and cognitive development. Early Hum Dev 31:181-193, 1993 48. Rogers B Feeding in infancy and later ability and attainment: A longitudinal study. Dev Med Child Neurol20421426,1978 49. Rosenblatt JS: Psychobiology of maternal behavior: Contribution to the clinical understanding of maternal behavior among humans. Acta Paedratri 397(suppl):?-S, 1994 50. Ryan A S The resurgence of breastfeeding in the United States. Pediatrics 99:12, 1997 51. Saigal S, Hoult LA, Streiner DL, et a1 school difficulties at adolescence in a regional cohort of children who were extremely low birth weight. Pediatrics 105:325-331,2000 52. Sommerfelt K, Ellertsen B, Markestad T: Low birthweight and neuromotor development: A population based, controlled study. Acta Paediatr 85604410,1996 53. Sung M,Vohr B, Oh W Growth and neurodevelopmental outcome of very low birth weight infants with intrauterine growth retardation: Comparison with control subjects matched by birth-weight and gestational age. J Pediatr 124:618-624,1993 54. Taylor B, Wadsworth J: Breast feeding and child development at five years. Dev Med Child Neurol26:73-80,1984 55. Temboury MC, Otero A, Polanco I, et al: Influence of breast-feeding on the infant’s intellectual development. J Pediatr Gastroenterol Nutr 1832-36,1994 et al: Developmental outcome of very low 56. Thompson RJ, Goldstein RF, &Her JM, birth weight infants as a function of biological risk and psychosocial risk. J Dev Behav Pediatr 15232-238, 1994 57. Uauy RD, Birch DG, Birch EE, et a1 Effects of dietary omega-3 fatty acids on retinal function of very-low-birth-weight neonates. Pediatr Res 28485492,1990 58. Weisglas-Kuperus N, Baerts W, Smrkovskey M, et al: Effects of biological and social factors on the cognitive development of very low birth weight children. Pediatrics 92:658-665, 1993
Address reprint requests to Ann Reynolds, MD Child Development Unit B-140 The Children’s Hospital University of Colorado Health Sciences Center 1056 E 19th Avenue Denver, CO 80218
0031-3955/01 $15.00 + .OO
BREASTFEEDING AND BRAIN DEVELOPMENT Ann Reynolds, MD
Although children who are breastfed have better neurodevelopmental outcomes, whether this is a biological or nutritional effect, an environmental effect, a genetic effect, or some combination of these factors is unclear. Because a randomized trial comparing breastmilk with formula is impossible, investigators must rely on observational studies, which creates methodologic difficulties. Potential confounding variables must be identified and addressed. Women who choose to breastfeed tend to have attributes, such as a higher level of education, that are associated with better developmental outcome12;however, evidence also supports a biological or nutritional role. Docosahexaenoic acid (DHA), a longchain polyunsaturated fatty acid present in large quantities in the brain and retina, is present in human milk but not in commercial formula. The role of human milk in neurodevelopmental outcome, with emphasis on visual and cognitive outcome in term and preterm infants, is reviewed. POLYUNSATURATED FAlTY ACIDS IN THE BRAIN
The role of polyunsaturated fatty acids in the function of the brain and retina is an area of active research. Although DHA and arachidonic acid are not present in infant formula, they can be synthesized from their essential fatty acid precursors a-linolenic acid (ALA) and linoleic acid, which are present in infant formula; however, an extra step seems to exist in the synthesis of DHA that slows down the conversion of ALA ~
From the Child Development Unit, The Children's Hospital, University of Colorado Health Sciences Center, Denver, Colorado PEDIATRIC CLINICS OF NORTH AMERICA VOLUME 48 NUMBER 1 FEBRUARY 2001
159
160
REYNOLDS
to DHA." This step is potentiallyProblematic in infants who are rapidly incorporating DHA into their CNS. Without a dietary source, infants must synthesize large amounts of DHA and arachidonic acid to keep up with the needs of the developing brain. The most rapid accretion of DHA and arachidonic acid in the human brain occurs during the last trimester of pregnancys when these long-chain polyunsaturated fatty acids normally would be provided by the placenta." Rapid accretion of DHA during the last 3 months of gestation would make premature infants vulnerable to a deficiency of DHA, but evidence shows that dietary DHA also may affect term infants. In a study35of infants who died of sudden infant death syndrome, DHA concentrations in the cortex were initially similar in newborns whether they received human milk or infant formula (7%); however, during the first 48 weeks of life, DHA concentrations in the cortex increased in infants who were breastfed to lo%, whereas concentrations in formula-fed infants remained the same. In another study of term infants who died of SIDS,'6 a postmortem examination found greater parietal cortex DHA in breastfed infants (9.7%) than in formula-fed infants (7.6%) at 9 weeks of age (P < 0.02). Infants who receive DHA in their diets seem to incorporate more DHA into their brains; however, whether any functional advantage to having higher levels of DHA in the brain exists is unclear. VISUAL FUNCTION IN BREASTFED INFANTS
Photoreceptor membranes in the retina have the highest concentration of DHA in the body. Photoreceptor cells undergo rapid maturation in the last trimester of pregnancy and the first 4 to 6 months postpartum. Consequently, investigators speculate that dietary availability of DHA should correlate with visual function. Neuringer et ala found functional visual deficits in rhesus monkeys who were deficient in DHA and ALA. Many studies also have been done in human infants to determine the effects of human milk or dietary fatty acids on visual function, the results of which have been incon~istent.'~ A positive effect of dietary DHA on visual outcome is more likely to occur in premature infants than in term infant^.^, 19,34, 57 Uauy et a157compared breastfed premature infants with three groups of formula-fed infants. The formula-fed infants were randomly assigned to formula deficient in 0-3 fatty acids, standard formula, or formula supplemented with marine oil that contains the w3 long-chain polyunsaturated fatty acids DHA and eicosapentaenoic acid. Thresholds for rod electroretinograms (ERGs) were significantly higher, and amplitudes of rod ERGs were significantly lower for the deficient group compared with the marine-oil-supplemented and breastfed groups. Results for the infants supplemented with marine oil were similar to those of breastfed infants. The infants fed standard formula had intermediate results that did not differ sigruficantly from the deficient groups or from the supplemented and breastfed gr0ups.5~The deficient group was deficient not only in DHA but also in its precursor, ALA. Carlson et a17 studied 67 premature infants randomized to stan-
BREASTFEEDING AND BRAIN DEVELOPMENT
161
dard formula or marine-oil-supplemented formula. Visual acuity was measured by preferential looking, a behavioral measure that can be used for infants. Infants prefer to look at a pattern instead of a homogeneous background; if an infant prefers a stimulus card with a checkerboard to a blank card, then one can assume that the squares in the checkerboard are larger than the infant’s threshold of resolution.41At 2 and 4 months of age, visual acuity measured by preferential looking was better in the marine-oil-supplemented group. No significant difference was found between groups at 6.5, 9.0, and 12.0 months. Visual acuity seems to ”catch up” in infants fed standard formula. It may have been more difficult to find a difference in the older age groups because visual acuity changes quickly in the first 6 months of life and then plateau^.^ In the same group at 9 and 12 months, Carlson6 reported increased discrete looks to faces during the novel portion of testing in the marine-oilsupplemented infants, which may indicate a “higher order of mental function.” Birch et a13also found better stereo acuity and letter matching in breastfed infants at 36 months of age. Most studies that have evaluated visual function did not adjust for socioeconomic status (SES), but a study by Makrides et a P showed that SES seemed to be associated with visual function. Another problem in interpreting these studies is that many of them use different oils for supplementing formula, many of which contain different types and.amounts of fatty acids, so comparison of the studies is difficult. Many of the studies also have small sample sizes that may limit their power.37The effect of breastfeeding on visual function is still unclear, especially in term infants. In preterm infants, there may be enhanced maturation of the visual system, which may affect the quantity and quality of early stimuli. Also, breastfeeding may affect other aspects of visual function that are seDarate from visual acuity. Future &dies should include larger sampfe sizes and adjust for SES. DEVELOPMENTAL AND COGNITIVE OUTCOME OF BREASTFED INFANTS The effect of breastfeeding on later neurodevelopmental outcome is controversial. Because physicians must rely on observational studies, unforeseen biases may exist in the samples and may affect the validity of the results. Confounding variables also make the results difficult to interpret. Potential confounding variables include: Parental or environmental variables Maternal and paternal IQ (as it relates to genotype and environment) Maternal and paternal education level SES Maternal age Home environment Mother-infant interactions and bonding
Hormonal influences of lactation on mother Maternal stress Maternal depression Maternal attitudes regarding health practices Birth order Child variables Neonatal factors that might affect decision to breastfeed, such as illness or feeding dysfunction Birth weight Intrauterine growth retardation Gestational age Abnormal neurologic examination Abnormal neuroimaging Bronchopulmonary dysplasia Meningitis or sepsis Skin-to-skin contact in premature infants Although this list is long, it may be incomplete. Women who breastfeed have different attributes from those who choose not to breastfeed. Women who choose to breastfeed tend to be older, have more education, 5o These factors are associated with a better and have higher developmental outcome for Lucas et-a132noted that women who were married, primiparous, delivered by cesarean section, and had a boy were most likely to initiate breastfeeding in preterm Other differences may be more difficult to measure among mothers who choose to breastfeed and mothers who do not.= In a study9of mothers of low SES, mothers who chose to breastfeed had better vocabulary, were older, and were less likely to smoke cigarettes. The mothers who breastfed also were found to be less authoritarian, and they provided a more stimulating home environment. Another study"" found that mothers who breastfeed may be more likely to practice health-enhancing behaviors, such as not smoking, eating five or more fruits and vegetables per day, and visiting the dentist yearly. These health-enhancing behaviors may have multiple implications, including a better nutritional state during pregnancy. In addition to the prelactation differences among women who choose to breastfeed and mothers who do not, breastfeeding may affect mothers, infants, or both. Breastfeeding is a time of mother-infant bonding. Early interactions between mothers and infants seem to help to determine the type of relationship that infants will have with their environments over time.12.22 The hormonal influences of breastfeeding also may affect maternal behavior. In rats, oxytocin has been found to trigger maternal behavior in estrogen-primed In humans, it is more difficult to differentiate learned from hormonally driven behavior. In rats and sheep, hormonal influences on maternal behavior seem to wane after delivery. Nonhormonal, or psychological, influences take over as rats learn the odor or call of their pups.49Even if hormonal influences are present early on in humans, they may not be critical in
BREASTFEEDING AND BRAIN DEVELOPMENT
163
influencing maternal behavior. Maternal stress and depression also can affect infant development?, 56 Kavanaugh et aIz6reported emotional rewards in mothers who provided breastmilk for their premature infants. This greater sense of well-being in mothers also may affect infants. Many of these variables could be addressed by evaluating environmental stimulation and mother-hild interactions. Few studies have attempted to do that in a systematic way.E,39 Several studies have evaluated developmental or cognitive outcome in breastfed infants.* A meta-analysis2 was published in 1999. To be included in the meta-analysis, a study had to compare subjects who were predominantly breastfed with subjects who were predominantly formula-fed and use an outcome measure that was a ”widely applied test of cognitive development.” Twenty studies were used, but only 11 provided usable data with adjusted and unadjusted scores. All of the studies included in the meta-analysis adjusted for at least five confounding variables. The analysis combined studies of term and preterm infants. After adjusting for possible confounding variables, an average 3.2point higher ”cognitive development score” was found among breastfed infants. Figure 1 shows the covariate-adjusted mean differences, with confidence intervals, for the studies included in the meta-analysis. The advantage was seen early on (6-23 mo) and continued throughout childhood to 10 to 15 years of age? If the cognitive advantage of breastfeeding were entirely caused by environmental factors, then the difference in developmental scores would be expected to increase with age. Developmental evaluation tools primarily evaluate sensorimotor skills in the first 2 or 3 years of life and therefore are less affected by the environment. As children get older, the tools incorporate more language, which is heavily influenced by the The meta-analysis also found that duration of breastfeeding correlated with developmental and cognitive outcome. No significant difference was found when duration of breastfeeding was 4 to 7 weeks. The mean difference increased as the duration of breastfeeding increased. The difference was 2.91 among infants who were breastfed for 28 weeks or more.2As with studies of visual function, a bigger advantage is conveyed by providing human milk to premature infants compared with term infants. In the metaanalysis, low birth weight infants had a 5.18-point advantage if breastfed compared with a 2.66-point advantage among term infants.2At least one of the studies in the analysis had a major methodologic concern because breastfeeding was determined by record review of charts that were not always complete.18Two of the reports13,24 were letters to the editor in response to another article and so were not peer reviewed; however, one letter reported data from a study that was published laterE and is discussed in this article. Table 1 summarizes studies included in the meta-analysis. Three studies evaluating outcome in term infants fed breastmilk that were published after the meta-analysis are discussed here.”, E, 36 *References 1, 13, 17, 24, 27, 43,4548, 52, 54, and 55.
164
REYNOLDS
3.40
Rogers
229
Fergusson et al
5.61 Ounsted et al
4.36
Morley et al
4.72
Morrow-Tlucak et al
5.02
Doyle et al
Jacobson and Jacobson
1.71 8.30
Lucas et al
2.70
Rogan and Gladen
4.36
Ternboury et al
3.70
Florey et al
3.16
1
All studies - 4 - 2
0
2
4
6
8
1 0 1 2 1 4
Figure 1. Effect of breastfeedingversus formula feeding on cognitive developmentalscore: covariate-adjustedmean differences with confidence intervals. (from Anderson JA, Johnstone BM, Remley DT: Am J Clin Nutr 70:525-535,1999; with permission.)
Jacobson et a P reported the cognitive outcomes of 280 children at 4 and 11years of age. A sigdcant association was found between the duration of breastfeeding and IQ at 4 and 11 years of age. The association was weakened but remained significant after adjusting for SES and maternal education level. The association was not statistically significant after additional adjustment for maternal IQ and the score on the Home Observation for Measurement of the Environment (HOME)? The Peabody Picture Vocabulary Test-Revised14 was used to measure IQ. It is not a measure of IQ but of receptive vocabulary. It was used because it correlates with the Wechsler Adult Intelligence Scale (median’ r = 0.72). The HOME inventory combines observation and interview of the primary caregiver in the child’s home. The HOME inventory evaluates opportunity for stimulation and quality and amount of maternal interact i ~ nThe . ~ HOME inventory is ”sigzufrcantly and independently related to children’s cognitive development” even when mothers’ IQ, educational achievement, and SES are adjusted accordingly.2OA possible concern with this studyz is that the cohort was derived from a study of
BREASTFEEDING AND BRAIN DEVELOPMENT
165
Table 1. SUMMARY OF STUDIES USED IN META-ANALYSIS
Study
No. Patlents
Rogers"
2424
Fergusson"
1037
ounsteda
242
Morley"
771
Morrow-Tl~cak~~ 229
Doyle13
96
Jacobsonz4
323
Lucas31
283
~0gan47
855
Tembow Florey'S
229 592
Tests Used to Evaluate Outcome
Additional Information
Covarlate Adlusted Mean Difference
8 y: picture intelligence; 1946 cohort, tests 3.40' 15 y: nonverbal not adequately ability, sentence described completion 2.09' Controlled for 3 y: Peabody Picture Vocabulary Test; many things, 5 y: Stanford Binet; including home 7 y: WE€-R environment 7.5 y: British Abilities All mothers had 5.61* hypertension Scale (very large confidence interval) 18 mo: Bayley, Neuroimaging 4.36* Developmental not addressed, Profile II LBW 4.72' 6,18, and 24 mo: Difference only significant at 12 Bayley and 24 mo, controlled for home environment, low SES group 5.02' 2 y: Bayley; 5 y: WPPSI; Letter, LBW 8 y: WISC 1.71 4 y: McCarthy Letter, same cohort as Jacobson (1999) 7.5-8.0 y: WISC No discussion of 8.30' neuroimaging (Anglicized revised edition) or neurological status, LBW 2.70' 6,12,18,24mo: Bayley; Not always blinded, PCB 3,4,5 y: McCarthy cohort 18-29 mo: Bayley 4.36 18 mo: Bayley Breastfeeding 3.70' determined by retrospective record review
'Covariate adjusted mean differrnce is significant. WISC-R = Wechsler Intelligence Scale for C h i l b (Revised); Bayley = Bayley Scales of Infant Development; McCarthy = McCarthy Scales of Children's Abilities;LBW = Low birth weight; WPF'SI = Wechsler Preschool and Primary Scale of Jntelligence; SES = socioeconomicstatus; PCB = polychlorinated biphenyls.
long-term effects of polychlorinated biphenyls. To participate in the study, mothers had to have a history of consuming at least 11.8 kg of fish from Lake Michigan over a 6-year peri0d.2~This requirement may have caused some bias. The authors note that prenatal exposure to polychlorinated biphenyls can cause subtle cognitive deficits but exposure by human milk should not. Also, all of the potential confounding variables may be highly correlated with each other, which may create multicolinearity, making the coefficients obtained in a hierarchical regression possibly misleading.'O When maternal IQ and HOME score are added into the multiple-regression analysis, breastfeeding, SES, and education level are insignificant. The authors do not address colinearity; thus, to what extent maternal IQ and HOME score are related to breastfeeding is unknown. This question could be answered by looking at zero-order correlation. A stepwise, multiple regression adding in first HOME score then maternal IQ, SES, education level, and finally breastfeeding could be conducted. Then the extent to which breastfeeding added predictive value to the dependent outcome measures could be determined. Unfortunately, the tables in the article are not displayed in a traditional hierarchical manner that would have detailed the amount of variance in scores accounted for by each block of variables detailing incremental predictions for each block. This is a perfect example of how difficult it is to interpret data in a nonrandomized, observational study. This is especially true when confounding variables are highly correlated with each other. In another study initially designed to follow infants exposed to chloride-deficient soy formula between 1978 and 1979,36 outcome at 9 and 10 years of age was assessed. This study also showed no cognitive advantage for breastfed infants at 9 and 10 years of age after scores were adjusted for maternal education level, paternal education level, and household annual income. In this study, mothers were asked to retrospectively report the duration of breastfeeding at the 9-year or 10-year well-child visit.% This method is concerning because one study found that only 54% of mothers could correctly identify within 1 month when they had switched from human milk to formula when their children were 2 to 4 years of age.I5 The third study is a population-based study from a cohort in New Zealand who were followed up until 18 years of age. Breastfeeding status was determined prospectively. In that study, a 0.11 to 0.30 standard deviation greater score (correlates with 1.54.5 points) was found on IQ and achievement measures among individuals who had been breastfed for more than 8 months even after adjustment for confounding factors, including maternal age, maternal education level, SES, family income, and an informal assessment of the en~ironment.2~ All of these studies evaluated long-term outcome. Any biological effect may be picked up more easily early on, before environmental factors become predominate, but clinically, long-term outcome would be the endpoint. In term infants, the effect of breastfeeding on later cognitive outcome seems to be small or insignificant after adjusting for maternal and environmental
-
BREASTFEEDING AND BRAIN DEVELOPMENT
167
factors. Additional studies should focus on lower SES groups because one study found a 4.7-point advantage in lower SES children even after adjusting for the HOME score, parental authoritarianism, maternal Peabody Picture Vocabulary Test-Revised, smoking habits, and marital ~fatus.3~ Future studies might also focus on more specific areas of cognition and neurologic function. Neuropsychological testing might be able to detect more significant differences in areas such as memory, attention, and fine motor skills. PREMATURE INFANTS
Premature infants are discussed separately because their nutritional needs are different from those of term infants and because the effects of breastfeeding seem to be greater in premature infants. Most of the studies published on cognitive outcome in premature infants who are fed human milk present data from the same cohort of infants in the United Kingdom. The preterm infants ( 4 8 5 0 g) were followed prospectively as part of a study to determine the optimal type of nutrition in children who were not breastfed or whose mothers did not produce adequate breast milk. Comparisons were made between preterm formula and unfiltered donor breast milk, and between preterm formula and term formula.30,31, 33, 38 In this cohort, Lucas et aI3l compared preterm infants who received their mothers’ breast milk during their hospital course with infants who did not. The comparison group received preterm formula (51%),mature pasteurized donor drip milk without fortification (31%), or term formula (18%). Infants in the group fed their mothers’ milk were supplemented with preterm formula, donor breast milk, or term formula in the same ratios. Lucas et a131found a 10.9-point advantage in IQ measured by the Wechsler Intelligence Scale for Children (Revised) (WISC-R) at 7.5 to 8 years of age. After controlling for SES and maternal education level, an 8.3-point advantage remained. In an attempt to further control for maternal differences associated with choosing to breastfeed, the investigators compared those who chose to provide breast milk but were unable to with those who never intended to provide breast milk. The mean IQ of children whose mothers intended to provide breast milk but could not (94.8) was similar to the mean IQ of children whose mothers never intended to breastfeed (92.8). They were both sigruficantly different from the group who received their mothers’ milk (103.7). While in the nursery, neonates received breast milk principally by nasogastric tube. To control for the effects of suckling, Lucas et a131 excluded infants from the analysis who would be breastfed after discharge. The breast milk advantage remained but decreased slightly to 7.5 points. Because many infants who received their own mothers’ breast milk also required supplementation, Lucas et a131 compared the amount of mother’s milk received with IQ. A significant dose-response curve was found. Comparisons of the group who received mother’s milk with the group that did not showed no significant
168
REYNOLDS
difference in birth weight, gestational age, days in the study, days to enteral feeds, and the percentage Ventilated for more than 5 days. More boys were included in the group that received mother's milk (55% versus 42%) and, as expected, SES and maternal education level were higher in the group that provided breast milk. These data are compelling but should be interpreted with caution. The study by Lucas et aP' has two potential problems. The heterogeneity of feeding type in the comparison group makes difficult the generalization of these findings. Term formula is inadequate to meet the nutritional needs of premature infants, so how much its use affected the overall results is unclear. It would be helpful to see the individual relationships between the mothers' milk group and those fed with preterm formula or with those fed donor breastmilk. A significant potential problem with the study is that no mention was made of rates of intrauterine growth retardation, intraventricular hemorrhage, periventricular leukomalacia, ventricular dilitation, abnormal neurologic status, or severe infection?l All of these factors could have a sigruhcant effect on developmental outcome.4,40, 53 Some children in the cohort seemed to have intraventricular hemorrhage and cerebral palsy because those issues are addressed in other p~blications~~~ m; however, they are addressed as they apply to preterm formula compared with term formula or donor breast milk, not as they apply to mothers' milk compared with other feeding strategies. Additional study is needed to determine the role of breast milk in the cognitive outcome of premature'infants. Future studies should include more complete information regarding confounding variables, such as intraventricularhemorrhage, periventricular leukomalacia, Cerebral palsy, other neurologic dysfunction, maternal intelligence, and a measure of environmental stimulation. Maternal stress and depression and mother-infant bonding also would be important to address in preterm infants. Despite these concerns at the least, the type of enteral feeding should be considered a potential confounding variable in neurodevelopmental outcome studies in premature infants. SUMMARY
Although biochemical evidence seems to support the fact that more DHA is incorporated into the brain of breastfed infants compared with formula-fed infants, whether the levels of DHA in the brain are clinically significant is unclear. Because randomized trials cannot be done, this issue is difficult to study. The effects of breastfeeding on developmental outcome in term infants seems to be small or insignificant. For otherwise healthy children the potential differences are not clinically relevant; however, these small differences distributed over an entire population might have a significant effect on society. Although significant methodologic concerns exist, the effects of breastfeeding on preterm infants may be greater than those for term infants. Extremely low birth weight, premature infants (<750-1000 g) have been found to have IQs that are
BREASTFEEDING AND BRAIN DEVELOPMENT
169
13 points lower than term controls and a 50% to 60% risk for requiring special-education services when they are in school.2*, 51 In these infants, small improvements in IQ and neurologic function could have a much greater effect. Further study of neurodevelopmental outcome in premature infants fed breast milk compared with those fed preterm formula are indicated. This information should not change the practice of encouraging breastfeeding of term and preterm infants because other advantages to breastfeeding exist. References 1. Agostoni C, Trojan S, Bellu R, et a1 Developmental quotient at 24 months and fatty acid composition of diet in early infancy: A follow up study. Arch Dis Child 76:421424,1997 2. Anderson JW, Johnstone BM, Remley DT: Breastfeeding and cognitive development: A meta-analysis. Am J Clin Nutr 70525-535, 1999 3. Birch E, Birch D, Hoffman D, et a1 Breastfeeding and optimal visual development. J Pediatr Ophthalmol Strabismus 3033-38,1993 4. Brazy JE, Eckerman CO, Oehler JM, et a1 Nursery neurobiological risk score: Important factors in predicting outcome in very low birth weight infants. J Pediatr 118:783-792, 1991 5. Caldwell BM, Bradley RH: Home Observation for Measurement of the Environment. Little Rock, AR, University of Arkansas Press, 1979 6. Carlson S E Lipid requirements of very-low-birth-weight infants for optimal growth and development. In Dobbing J, Benson JD (eds): Lipids, Learning and the Brain: Fats in Infant Formulas. Report of the 103rd Ross Conference on Pediatric Research. Columbus, OH, Ross Laboratories, 1993 7. Carlson SE, Werkman SH, Rhodes PG, et a1 Visual-acuity development in healthy preterm infants: effect of marine41 supplementation. Am J Clin Nutr 58:35-42,1993 8. Clandinin MT, Chappell JE,Leong S, et al: Intrauterine fatty acid accretion rates in human brain: Implications for fatty acid requirements. Early Hum Dev 4:121-129,1980 9. Cogill SR, Caplan HL, Alexandra H, et al: Impact of maternal postnatal depression on cognitive development of young children. BMJ 292:1165-1167,1986 10. Cohen J, Cohen P: Multiple regression and correlation: Two or more independent variables. In Cohen J, Cohen P (eds): Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences. Hillsdale, NJ, Lawrence Erlbaum Associates, 1975 11. Crawford MA: Placental delivery of arachidonic and docosahexaenoic acids: Implications for the lipid nutrition of preterm infants. Am J Clin Nutr 71(supp1):275-284,2000 12. De Andraca I, Uauy R Breastfeeding for optimal mental development. The Alpha and the Omega in human milk. World Rev Nutr Diet 781-27,1995 13. Doyle LW, Rickards AL, Kelly EA, et a1 Breastfeeding and intelligence [letter]. Lancet 339744-745,1992 14. Dunn LM Peabody Picture Vocabulary Test-Revised. Circle Pines, MN, American Guidance Services, 1987 15. Eaton-Evans J, Dugdale AE: Recall by mothers of the birth weights and feeding of their children [abstract]. Human Nutr Appl Nutr 40371-175, 1986 16. Farfuharson J, Cockburn F, Patrick WA, et al: Infant cerebral cortex phospholipid fattyacid composition and diet. Lancet 340810-813,1992 17. Fergusson DM, Beautrais AL, Silva PA: Breastfeeding and cognitive development in the first seven years of life. Soc Sci Med 1617051708,1982 18. Florey CdV, Leech AM, Blackhall A Infant feeding and mental and motor development at 18 months of age in first born singletons. Int J Epidemiol24(suppl):21-26, 1995 19. Gibson RA, Makrides M Polyunsaturated fatty acids and infant visual development: A critical appraisal of randomized trials. Lipids 34A79-184, 1999
20. Gottfried AW. Home environment and early cognitive development:Integration, metaanalyses, and conclusions. In Gottfried AW (ed): Home Environment and Early Cognitive Development, Longitudinal Research. Orlando, Academic Press, 1984 21. Hack M, Taylor G, Klein N, et al: School-age outcomes in children with birth weights under 750 grams. N Engl J Med 331:753-759,1994 22. Hofer MA: Early relationships as regulators of infant physiology and behavior. Acta Paediatr 397(suppl):9-18,1994 23. Horwood LJ, Fergusson DM: Breastfeeding and later cognitive and academic outcomes. Pediatrics 101:91-97, 1998 24. Jacobson SW, Jacobson JL: Breastfeeding and intelligence [letter]. Lancet 339926, 1992 25. Jacobson SW, Chiodo LM, Jacobson J L Breastfeeding effects on intelligence in 4 and 11-year old children. Pediatrics 10371, 1999 26. Kavanaugh K, Meier P, Zimmerman B, et a1 The rewards outweigh the efforts: Breastfeeding outcomes for mothers of preterm infants. J Hum Lact 13:15-21,1997 27. Lanting CI, Fidler V, Huisman M, et al: Neurological differences between 9-year-old children fed breast-milk or formula-milk as babies. Lancet W1319-1322, 1994 28. Lawrence RA, Lawrence RM: Psychologic impact of breastfeeding. In Breastfeeding: A Guide for the Medical Profession, ed 5. St. Louis, Mosby, 1999 29. Lucas A, Morley R, Cole TJ: Randomised trial of early diet in preterm babies and later intelligence quotient. BMJ 3171481-1487, 1998 30. Lucas A, Morley R, Cole TJ, et a1 A randomised multicentre study of human milk versus formula and later development in preterm infants. Arch Dis Child 70F141F146,1994 31. Lucas A, Morley R, Cole TJ, et a1 Breast milk and subsequent intelligence quotient in children born preterm. Lancet 339261-264,1992 32. Lucas A, Cole TJ, Morley R, et a1 Factors associated with maternal choice to provide breastmilk for low birth weight infants. Arch Dis Child 63:48-52,1988 33. Lucas A, Gore SM, Cole TJ, et a1 Multicentre tial on feeding low birth weight infants: Effects of diet on early growth. Arch Dis Child 59:722-730, 1984 34. Makrides M, Neumann MA, Jeffrey B, et al: A randomized trial of different ratios of linoleic to a-linolenic acid in the diet of term infants: effects on visual function and growth. Am J Clin Nutr 71:120-129, 2000 35. Makrides M, Neumann MA, Byard RW, et a1 Fatty acid composition of brain, retina, and erythrocytes in breast- and formula-fed infants. Am J Clin Nutr 60189-194,1994 36. Malloy MH, Berendes H Does breast-feeding influence intelligence quotients at 9 and 10 years of age? Early Hum Dev 50:209-217,1998 37. Morley R Nutrition and cognitive development. Nutrition 14752-754,1998 38. Morley R, Cole TJ, Powell R, et al: Mother’s choice to provide breast milk and developmental outcome. Arch Dis Child 63:1382-1385,1988 39. Morrow-Tlucak M, Haude RH, et al: Breastfeeding and cognitive development in the first 2 years of life. SOCSci Med 26635-639, 1988 40. Msall ME, Buck GM, Rogers BT, et al: Risk factors for major neurodevelopmental impairments and need for special education resources in extremely premature infants. J Pediatr 119:606414, 1991 41. Neuringer M Infant vision and retinal function in studies of dietary long-chain polyunsaturated fatty acids: Methods, results, and implications. Am J Clin Nutr 71(suppl):256-267,2000 42. Neuringer M, Connor WE, Lin DS, et ak Biochemical and functional effects of prenatal and postnatal omega 3 fatty acid deficiency on retina and brain in rhesus monkeys. Proc Natl Acad Sci U S A 83:4021-4025,1986 43. Ounsted M, Moar V, Cockbum J, et al: Factors associated with intellectual ability of children born to women with high risk pregnancies. BMJ 2881038-1041,1984 44. Pesaa JA, Shelton MN: Health-enhancing behaviors correlated with breastfeeding among a national sample of mothers. Public Health Nurs 16:120-124,1999 45. Pollock J I Long-term associations with infant feeding in a clinically advantaged population of babies. Dev Med Child Neurol x429-440, 1994 46. Richards M, Wadsworth M, Rahimi-Foroushani A, et ak Infant nutrition and cognitive
BREASTFEEDING AND BRAIN DEVELOPMENT
171
development in the first offspring of a national UK birth cohort. Dev Med Child Neur 40:16%167,1998 47. Rogan WJ, Gladen BC: Breast-feeding and cognitive development. Early Hum Dev 31:181-193, 1993 48. Rogers B Feeding in infancy and later ability and attainment: A longitudinal study. Dev Med Child Neurol20421426,1978 49. Rosenblatt JS: Psychobiology of maternal behavior: Contribution to the clinical understanding of maternal behavior among humans. Acta Paedratri 397(suppl):?-S, 1994 50. Ryan A S The resurgence of breastfeeding in the United States. Pediatrics 99:12, 1997 51. Saigal S, Hoult LA, Streiner DL, et a1 school difficulties at adolescence in a regional cohort of children who were extremely low birth weight. Pediatrics 105:325-331,2000 52. Sommerfelt K, Ellertsen B, Markestad T: Low birthweight and neuromotor development: A population based, controlled study. Acta Paediatr 85604410,1996 53. Sung M,Vohr B, Oh W Growth and neurodevelopmental outcome of very low birth weight infants with intrauterine growth retardation: Comparison with control subjects matched by birth-weight and gestational age. J Pediatr 124:618-624,1993 54. Taylor B, Wadsworth J: Breast feeding and child development at five years. Dev Med Child Neurol26:73-80,1984 55. Temboury MC, Otero A, Polanco I, et al: Influence of breast-feeding on the infant’s intellectual development. J Pediatr Gastroenterol Nutr 1832-36,1994 et al: Developmental outcome of very low 56. Thompson RJ, Goldstein RF, &Her JM, birth weight infants as a function of biological risk and psychosocial risk. J Dev Behav Pediatr 15232-238, 1994 57. Uauy RD, Birch DG, Birch EE, et a1 Effects of dietary omega-3 fatty acids on retinal function of very-low-birth-weight neonates. Pediatr Res 28485492,1990 58. Weisglas-Kuperus N, Baerts W, Smrkovskey M, et al: Effects of biological and social factors on the cognitive development of very low birth weight children. Pediatrics 92:658-665, 1993
Address reprint requests to Ann Reynolds, MD Child Development Unit B-140 The Children’s Hospital University of Colorado Health Sciences Center 1056 E 19th Avenue Denver, CO 80218